Pneumocystis jiroveci (Pneumocystis carinii) is an important cause of pneumonia (PCP) in immunocompromised individuals. The standard approach for diagnosing PCP is a microscopic examination of smears prepared from induced sputum or bronchial alveolar lavage (BAL) samples. Recently, investigators have been designing and testing PCR assays for the detection of P. jiroveci in respiratory samples. Of particular interest is the use of PCR with oral wash samples as a means of detecting P. jiroveci in the respiratory tract. These noninvasive specimens could prove to be of value for use in screening tests to rule out PCP. Microscopic methods are too insensitive to be useful with oral wash samples. The increased sensitivity of the PCR method, however, generates some positive results with samples obtained from patients who are only colonized or infected at a sub-clinical level. A precise quantitative method could help differentiate low level colonization from infection and, consequently, improve the clinical usefulness of PCR performed on oral washes and other respiratory samples. We have developed a rapid quantitative real time PCR assay targeting the MSG genes of P. jiroveci using fluorescence resonance energy transfer (FRET) detection probes for signal detection. A blinded, prospective study has been conducted with collaborators at UCSF to further evaluate the performance of the real time PCR assay in detecting PCP. For samples obtained within one day of the initiation of PCP therapy the sensitivity of oral wash samples with QTD-PCR for detecting PCP was greater than 90%. The results of this study are in press. A second collaborative study with UCSF is underway to attempt to detect P. jiroveci colonization of the upper airways of health care workers after exposure to patients with PCP. Recent work on an improved method for sample processing has demonstrated that use of DTT markedly improves the homogeneity of the processed oral wash material. If, as expected, this results in more reproducable quantitative results the positive and negative predictive values for PCP associated with certain levels of signal should improve considerably. We are initiating testing of a series of samples with the processing modifications to test this hypothesis.